KR20180032628A - Method and apparatus for detecting voltage in a supply network - Google Patents

Abstract

The present invention relates to a method of detecting a voltage in an electrical three-phase supply network according to amplitude and phase for fundamental oscillation and at least one harmonic, comprising the steps of: measuring the electrical three-phase voltage of the supply network ; Converting the measured voltage value to a polar coordinate including a rotation voltage indicator for the fundamental vibration as a measured reference indicator; Monitoring a value of each of the at least one voltage indicator for the fundamental vibration and the at least one voltage indicator for the at least one harmonic to be detected using the status monitor; And updating the monitored values according to the measured reference indicator.

Description

Method and apparatus for detecting voltage in a supply network

The present invention relates to a method of detecting a voltage in an electrical three-phase supply network according to amplitude and phase for fundamental oscillation and at least one harmonic. The present invention also relates in particular to a method for compensating for harmonics in a power supply network based on detection in accordance with the invention. The present invention also relates to a device for detecting voltage and an apparatus for harmonic compensation. The present invention also relates to a wind turbine in which such a method or a plurality of such methods are implemented or in which such a device or a plurality of such devices are installed.

In order to supply power to a three-phase electricity supply network in general, it is necessary to detect the voltage according to the amplitude and the phase, which also includes the detection or recognition of the frequency. The demand for detecting such voltages is increasingly increasing, which is also related to the speed for detecting changes, in particular. A method for quickly and accurately detecting a voltage in accordance with amplitude and phase is described, for example, in U.S. Patent No. 8,981,755.

In addition, wind turbines operating with the concept of a full converter, that is, wind turbines in which the power of a wind turbine is fully supplied to the network via an inverter, are able to adjust the supplied power to the target in the network with asymmetric qualities.

An example of the supply of an asymmetric three-phase current is described in US application US 2013/0141951. Here, a method of symmetrical components is used. Where the current components of the normal sequence system and the current components of the reversed phase sequence system are generated independently of each other and then superimposed and supplied with the desired unbalanced current.

The quality of this power supply also ultimately depends on the quality of the asymmetric detection. It should be noted that the concept of a supply network so far is that a large power plant with a dominant synchronous generator directly coupled to the network basically functions to determine the quality of the voltage, and in particular the symmetry, by the dominance of the current of the synchronous generator. For example, as described in U.S. Patent No. 6,924,627, although a primary consideration for compensating for asymmetry has already existed for a long time, the synchronous generator of the large power plant described above does not compensate for this compensation by active control, , The principles described above were nevertheless maintained and, ultimately, not so.

In modern energy systems, the generating unit and the consuming device are increasingly being connected to the transmission or distribution network via frequency converters or rectifiers. Due to the switching components of these power electrical systems, a sinusoidal 50 Hz or 60 Hz AC voltage is contaminated by harmonics in the frequency range up to about 10 kHz. These harmonics adversely affect the various operating means of the energy system, so if the level is too high, they are attenuated by the active or passive filter.

Wind turbines provide a significant portion of the energy already generated in some energy systems. Often, partly, wind turbines are connected to the network via a frequency converter, and thus, in the worst case, they contribute to the harmonics of the network as well.

The German Patent and Trademark Office has proposed in the priority application for this PCT application the following techniques: DE 100 32 447 A1, US 2012/0150468 A1, US 2014/0307488 A1, EP 2 223 405 B1, Detection of Fundamental and Harmonic Components of Grid Voltages Using Kalman Filter and Generalized Averaging Method "Akzabageri et al., &Quot; Extended Kalman Filter Based Grid Synchronization in the Presence of Voltage Unbalance for Smart Grids" A Kalman filtering approach to short-time Fourier analysis "(Robert Al.

It is therefore an object of the present invention to solve at least one of these problems mentioned, and in particular to suggest improvements. In particular, a solution should be proposed that creates the possibility of network improvement, including harmonic compensation of the network. At least alternative solutions to the known solutions should be proposed.

According to the invention, a method according to claim 1 is proposed. According to this, a voltage is detected in the electrical three-phase supply network according to the fundamental oscillation and amplitude and phase for at least one harmonic. Here, the electricity supply network should be understood as an electric distribution network, for example. This relates basically to an electric network to which many current generating devices and current consumption devices are connected.

First, the electrical three-phase voltage of the supply network is measured.

The measured electrical three-phase voltage is then converted to polar coordinates to generate a rotating voltage indicator for the fundamental oscillation. This is referred to as the measured reference indicator and is used correspondingly later. The conversion can be performed, for example, as disclosed in U.S. Patent No. 8,981,755.

Also with the status monitor, the value of each of the at least one voltage indicator for the fundamental vibration and the at least one voltage indicator for the detected harmonics is monitored. This refers to the status monitor in terms of control technology. The monitored value is now updated according to the measured reference indicator.

That is, a monitor model is based on at least one fundamental oscillation and at least one harmonic for displaying the network voltage, each using at least one voltage indicator. The measured system is represented by a voltage indicator for the fundamental oscillation, where through continuous measurement and conversion these voltage indicators may be varied in amplitude and phase.

The comparison for the updating of the monitors is carried out in particular by comparing these measured values in the polar coordinates, i. E. The values of this rotational voltage indicator by measurement, with the sum of all the voltage indicators based on the monitor.

Preferably, the status monitor operates according to the detected commercial frequency, and the detected commercial frequency is input as an input variable to the status monitor. Therefore, it is proposed that the status monitor does not monitor the commercial frequency as a state to be updated as needed for monitoring, but rather accepts and provides the value for the detected and inputted frequency as given. Stabilization of monitoring can be achieved by improving stability, or otherwise expressing it, because each of these frequencies, which is a specific variable of the voltage indicator, affects the rotation monitoring voltage indicator and represents a fixed variable. The monitor can focus on amplitude and phase in monitoring, and this can eliminate potential interferences between frequency and amplitude and phase on the one hand, or even threats to stability.

According to one embodiment, it is proposed that the status monitor monitors the value of the voltage indicator using the current monitor matrix, and the current monitor matrix is selected from a plurality of pre-computed monitor matrices according to the detected commercial frequency. It should be noted here that the terms matrix and matrix are used as synonyms. This monitor matrix represents a known component of the monitor structure and basically has the task of updating the monitored state based on the comparative measurements. Here, the reference indicator is used as the comparative measurement value. Here, in order to consider the variable frequency in monitoring, the calculation of the frequency dependent monitor matrix can be very costly. This cost may, depending on the circumstances, exceed the available computational capacity of the microcontroller at any given time step, so that calculation at the required clock rate is not possible. Depending on the frequency, this cost can be reduced, especially using a previously calculated monitor matrix selected by each currently detected frequency. Sometimes this allows for considerable consideration of the frequency change. At frequencies where there is no predetermined monitor matrix that is subject to the accuracy of frequency detection, this can be determined by interpolation between the two monitor matrices.

Preferably, a Kalman filter is used as the state monitor. In this regard, monitoring of the state may be referred to as filtering or the state monitor may be referred to as a filter. By the use of a Kalman filter, monitoring or filtering is particularly suitable for considering interference, especially measurement noise, i.e. filtering out.

According to one embodiment, it is proposed that for a DC component and a fundamental oscillation of a voltage to be detected and for each harmonic detected, a normal sequence system and a reverse phase sequence system are detected by the status monitor, respectively. This allows the asymmetry of the three-phase voltage to be detected by the state monitor as well. Although there is no serious asymmetry that does not need to be known before measurement, this detection of a reverse phase sequence system can also be performed. If desired, the components of the reverse phase sequence system are correspondingly small. Thus, in this detection of a normal sequence system and a reverse sequence system, monitoring is based on a method of symmetric components.

Preferably, two voltage indicators are respectively based on the voltage indicator for the DC component, the fundamental oscillation and each harmonic being detected, and the value is monitored, so that the voltage for both the voltage indicator and the reversed phase sequence system for the normal sequence system Indicator is adopted. Also for the DC component, it may be useful to adopt two values, the real part and the imaginary part of the voltage indicator, since the indicator can also be computed at least mathematically at the application of the method of the symmetric component here. Preferably, a very large number of harmonics, for example 25 harmonics, are detected, i.e. up to the 25th harmonic. In this case, if the first harmonic is the fundamental vibration, there are 51 indicators and therefore 102 values.

According to one embodiment, a method of detecting a voltage is characterized in that the measurement of the electrical three-phase voltage of the supply network is often done in a transformer connected to an electricity supply network, also referred to as a transformer. Can be measured in a simple and reliable manner in such a transformer, in particular in a transformer that allows power to be supplied to the electricity supply network.

Measurements in these transformers are particularly advantageous when power is supplied via such a transformer and the supply device considers measurement or detection of the voltage of the electricity supply network. Distortion can be prevented here, particularly when a current that is at least partially compensated for by the monitored harmonics is supplied through this transformer.

Measurements are preferably made on the side of the transformer electrically facing the wind turbine that supplies power to the electricity supply network through such a transformer. In particular, such a wind turbine can account for such detected voltage values, in particular monitored harmonics, and correspondingly adjust the supply of power. Especially to compensate for, or at least reduce, the monitored harmonics of the network voltage.

Preferably, the step of converting the measured voltage to polar coordinates is performed using Clarke-transformation under the adoption of zero sequence system degrees of freedom of voltage. This also corresponds to the transformation described in the above-mentioned U.S. Published Unexamined Patent Application (US 8,981,755). It is assumed here that there is no zero sequence system, which is also often employed in the application of the theory of symmetric components. However, the adoption of zero-sequence system degrees of freedom does not necessarily include the adoption of the degrees of freedom of DC components. Rather, since there may be various DC components per phase in a three-phase current network, they are also found in normal sequence systems and reverse sequence system components after conversion.

According to the present invention, a method of supplying power according to a corresponding independent claim, namely claim 9, is proposed. According to this, power is supplied to the electrical three-phase supply network,

Detecting a voltage in the supply network by way of at least one of the preceding embodiments (at least one harmonic is detected), and

Generating power to supply to the electricity supply network, the power including a compensating current component to reduce at least one detected harmonic;

.

This method therefore also presupposes an electrical three-phase supply network based on the above-described method for detecting voltage in such an electrical three-phase supply network. Here too, an electrical distribution network may be included for illustrative purposes only.

Therefore, the voltage is also detected by the above-described method. The method also includes correspondingly detecting, i.e. monitoring, the value of at least one harmonic of the voltage. Thus, there is a value according to the amplitude and phase for this harmonic, and it can be used to at least reduce and ideally completely compensate for this detected harmonic. The power supplied for this includes a compensating current component, which accordingly comprises at least one harmonic. That is to supply the corresponding power to the electricity supply network, in particular the current comprises a fundamental component or a fundamental oscillation. This basic oscillation, i. E. The ideal graph, is now also superimposed, i. E. In the electronic sense, the compensating current components. This compensating current component should reduce the detected harmonics or at least a part thereof.

According to one embodiment, it is proposed that the compensating current component comprises at least one current harmonic and the current harmonics are determined, respectively, via amplitude and phase and according to the monitored harmonics of the detected voltage. Hence monitoring the harmonics of the voltage provides information, ie a clearly defined description of the harmonics or harmonics of this voltage. The compensating current component is preferably similarly constructed here. That is, the compensating current component includes, for example, the corresponding current harmonics for the detected voltage harmonics. The amplitude of the current harmonics may be present or at least related to the harmonics of the voltage at a predetermined rate. Also, the phase is adjusted accordingly to compensate or at least reduce it.

Preferably, upon detection of a voltage, a normal sequence system and a reverse sequence system are detected for each harmonic. Preferably, correspondingly corresponding current harmonics also include a normal sequence system and a reverse sequence system or a normal sequence system component and a reverse sequence sequence system component. Here, the amplitude and phase for the current harmonic may also be determined, respectively. However, it is also possible that only the normal sequence system components are taken into consideration, and in particular that the compensating current components basically determine only one component for each current harmonic, and that the asymmetry which is based on observations of the symmetrical components at this point is ignored . The background to this is that observation of normal and reverse sequence systems for monitors may be useful to ensure overall functionality and overall quality and accuracy. However, it may often be sufficient to generate at power supply only the normal sequence system components.

Note that harmonics also refer to harmonics of the order as current harmonics, and a plurality of harmonics refers to harmonics each having a different frequency.

A further embodiment suggests that the amplitudes of the current harmonics at the compensating current components are adjusted through the regulator, respectively. The regulator receives here an adjustment difference between the voltage setpoint and the monitored voltage actual value of the harmonics of the detected voltage as an input signal. In the simplest case, this means that the voltage setting has a value of 0, and ultimately the harmonic is reduced to zero. Of course, it may be important not to set the value to 0 in certain cases. This difference value, also referred to herein as the regulation difference, is provided to the regulator between the voltage setpoint and the voltage actual value of the harmonic. This may be, for example, a PID-regulating device, a PI-regulating device or a PD-regulating device. In the simplest case, a P-regulator is used, which derives the amplitude of the current harmonic corresponding to this factor, taking into account the different units between voltage and current as well as the detected voltage difference. This allows the corresponding harmonics of the voltage in a pure P-regulator to be almost completely compensated.

To fully compensate for this harmonics of the voltage, for example a PID-regulating device is proposed. However, it should be noted that sufficient regulatory energies must be ensured in advance. In this case, it may be successful to fully or nearly completely compensate for the harmonics of the voltage by this PID-regulating device. If the regulatory energy is not sufficient, the regulator may reach its limit.

In some cases, the PD-regulator can generate the dynamics without requiring complete compensation.

According to the present invention, a corresponding claim, that is, an inverter according to claim 14, which produces power to be supplied to a three-phase power network is proposed. Such an inverter is prepared to operate in the manner described in accordance with at least one of the embodiments of the method for supplying power. In this regard, it is essentially a control device in which such a method for power supply, i. E. A method of generating the supplied current, is implemented.

Also, a wind turbine including such an inverter is proposed. The use of a wind turbine not only allows energy to be supplied to the electricity supply network, but also compensates or reduces the harmonics of the voltage, since such current can be supplied. Correspondingly or alternatively, it is proposed that the wind turbine also implement at least one method as described above, namely a method of detecting the voltage as well as a method of generating the current to be supplied.

The method described above may also be used to determine the harmonics of the network voltage as well as to determine the harmonics of the power. This is also proposed in accordance with the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The invention will now be described in more detail with reference to the accompanying drawings in accordance with an exemplary embodiment.

Figure 1 schematically shows a perspective view of a wind turbine.
2 shows a simplified structure for explaining a voltage detection method according to an embodiment.
Figure 3 shows the functional block diagram of Figure 2 in more detail.
Figure 4 schematically illustrates an apparatus for powering an electrical supply network by a wind turbine according to an embodiment of the invention.

Figure 1 shows a wind turbine 100 with a tower 102 and a nacelle 104. [ A rotor 106 having three rotor blades 108 and a spinner 110 is disposed in the nacelle 104. In operation, the rotor 106 is rotationally driven by the wind, thereby driving the generator in the nacelle 104.

로 지칭되는 극좌표로 변환한다.Figure 2 shows a measuring device 201 for detecting a voltage in an electrical three-phase supply network 202, schematically shown. Thus, the voltage U ₁ , ₂ , 3 is measured by the measurement sensor 204 in three phases. The measurement is performed especially for each sampling step of the subsequent evaluation. The measured voltage is supplied to a conversion block 206, which converts each of these three measured voltage values

Lt; / RTI >

The measured voltage is also input to the frequency determining block 208, which determines the frequency f, i.e., the commercial frequency of the supply network 202, from the measured voltage values. The determination of the frequency is thus made by this frequency determination block 208 through an algorithm that is independent of the Kalman filter. Here, for example, a filter for determining the commercial frequency from the spatial indicator voltage angle can be used. Preferably, alternatively, discrete Fourier transforms of three voltage values are used to determine the filtered angle of the network voltage. Again, the frequency signal can be extracted by appropriate filtering. The advantage of using angles obtained from the DFT, as compared to using the space indicator angle here, is the filter effect at integer harmonics of the fundamental frequency.

This block diagram of Figures 2 and 3 and in part Figure 4 serves to illustrate the method used. The individual functions of the illustrated blocks need not necessarily be performed in this separate block, but may be implemented together in a microprocessor or other computing unit.

) 및 검출된 상용 주파수(f)는 칼만 필터 블록(210)에 입력된다. 따라서, 칼만 필터 블록(210)은 검출된 상태(x)를 출력값으로서 결정하여 출력하는 필터이다. 여기서 이러한 상태 벡터(x)로 표현되는 이러한 검출된 상태는 특히 DC 성분과, 각각 기본 진동의 정상 시퀀스 시스템 및 역상 시퀀스 시스템과, 각각 검출된 고조파의 정상 시퀀스 시스템 및 역상 시퀀스 시스템에 대한 전압 인디케이터의 값들을 포함할 수 있다. 제1 차수가 기본 진동인 경우, 예를 들어 25차 고조파까지의 전압을 고려하면, 이러한 상태 벡터(x)는 102개의 값을 포함할 수 있다. 이러한 수는 DC 성분에 대한 2개의 값, 즉

및

로 지칭되는 실수부와 허수부에서 얻어진다. 기본 진동 및 각 고조파에 대해, 정상 시퀀스 시스템에 대해 각각 2개의 값 그리고 역상 시퀀스 시스템에 대해 각각 2개의 값, 즉 각각 인디케이터의 실수부 및 허수부가 얻어진다. 제1 차수로서, 기본 진동에 대해, 정상 시퀀스 시스템에 대한 값

및

그리고 이에 대응하여 역상 시퀀스 시스템에 대한 값

및

이다. 유사하게, 각각의 추가 차수에 대해서도 4개의 값이 또한 제공될 수도 있다. In each case, the converted polar coordinates (

And the detected commercial frequency f are input to the Kalman filter block 210. [ Therefore, the Kalman filter block 210 is a filter that determines the detected state ( x ) as an output value and outputs it. This detected state, represented here by this state vector ( x ), is in particular a DC component and a normal sequence system and a reversed phase sequence system, respectively, of the fundamental oscillation and a voltage indicator for the normal sequence system and the reverse phase sequence system, respectively, ≪ / RTI > If the first order is a fundamental, for example, taking into account the voltage up to the 25th harmonic, this state vector x may contain 102 values. These numbers are the two values for the DC component,

And

Lt; RTI ID = 0.0 > and imaginary < / RTI > For the fundamental oscillation and each harmonic, two values for the normal sequence system and two values for the reverse phase sequence system, respectively, i.e. the real and imaginary parts of the indicators, respectively, are obtained. As a first order, for the fundamental oscillation, the value for the normal sequence system

And

Correspondingly, the value for the reverse phase sequence system

And

to be. Similarly, for each additional order, four values may also be provided.

This value, i.e., the state vector thus detected, can be used to generate power to power the supply network 202.

) 및 주파수 결정 블록(208)에 의해 결정된 상용 주파수(f)를 입력 변수로서 수신한다. 이러한 특정 상용 주파수(f)에 의존하여 모니터 행렬(K)이 모니터 행렬 블록(302)에서 결정되고, 이에 따라 모니터 행렬[K(f)]로서 출력된다. 모니터 행렬 블록(302)은 주파수(f)에 따라 대응하는 모니터 행렬(K)을 테이블로부터 추출함으로써 이러한 모니터 행렬(K)을 결정한다. 필요에 따라, 모니터 행렬(K)이 정확한 주파수에 매칭되지 않으면, 보간에 의해 인접한 주파수의 모니터 행렬로부터 결정될 수 있다. The details of the Kalman filter block 210 are schematically illustrated in FIG. 3 illustrates this Kalman filter block 210 and, as already seen in FIG. 2, the measured and polar transformed voltage (< RTI ID = 0.0 >

And a commercial frequency f determined by the frequency determining block 208 as input variables. The monitor matrix K is determined in the monitor matrix block 302 and is output as a monitor matrix K (f), depending on this specific commutation frequency f. The monitor matrix block 302 determines this monitor matrix K by extracting the corresponding monitor matrix K from the table according to the frequency f. If necessary, the monitor matrix K can be determined from the monitor matrix of adjacent frequencies by interpolation, unless it matches the correct frequency.

Depending on the frequency, a system matrix [A (f)] is also determined in the system matrix block 304. In this case, the system matrix is calculated in the system matrix block 304 in a known manner.

)이 각각 고려된다. The system state x to be detected is basically continuously updated in the update block 306. [ Where such an update block 306 operates as a roughly system-defined model, where the current measured and converted to polar coordinates for updating or balancing (e.g.,

Respectively.

This also requires a monitor matrix [K (f)] for which the update block 306 always actually receives the monitor matrix from monitor matrix block 302 for this purpose. A system limit matrix C is also required and is therefore illustratively shown as an input to the update block 306. [

It also requires an updated state vector [ x (k + 1)]. This updated state vector x (k + 1) is computed in the state block 308 from the final state vector x (k) and the system matrix A (f). To this end, state block 308 requires a system matrix [A (f)] that actually each receives from system matrix block 304. State block 308 actually receives the previous state vector [ x (k)] from update block 306. [

Thus, the updated state vector [ x (k + 1)] is also applied as an input variable to update block 306 as described above. In particular, in order to start the process, initialization may be performed at initialization block 310 for this updated state vector [ x (k + 1)].

Thus, the Kalman filter 210 can now output each detected state vector x (k) as a state vector x . Optionally, this may also only be converted to a multiplication by a constant factor in the transform block 312 if necessary. Of course, in this case, the state vector x shown in FIG. 3 will be correspondingly different from the internal state vector [ x (k)].

This state vector x can now include values for the DC component, the indicator of the fundamental oscillation and the harmonic indicator, which can be used to determine the current to be supplied. This is shown in FIG. Although FIG. 4 shows this for wind turbine 400, other power supplies can also record these measurements and use them for current control, i.e., control of the current to be supplied.

The structure of FIG. 4 illustrates a supply network 402 powered by a transformer 404. Further, the wind turbine 400 generates a current, which is rectified by the rectifier 408 and can be supplied to the inverter 410, by the generator 406. Such an inverter 410 is also often referred to as inverter 410.

This inverter 410 generates a three-phase current (i _1,2,3 ) that can be supplied to the supply network 402 via the transformer 404.

It is now proposed to detect the voltage and frequency with the measuring sensor 412, which is symbolically shown, and supply it to the measuring block 414. [ The measurement block 414 may correspond to the measurement device 201 of Fig. 2 together with the measurement sensor 412. Fig. In this case, frequency f will be determined in the measurement block 414 from the previously detected voltage, but for purposes of explanation, frequency f in Fig. 4 is also shown as an input variable to the measurement block 414.

The measurement block 414 then calculates the values for the DC components, the fundamental oscillations and the harmonics, in particular the state vector ( x ), which may include values for the normal sequence system and the reverse phase sequence system, As an output variable.

This value, i. E. The state vector x , may be input to the control block 416, which may include, for example, a controller. In this control block 416, according to the detected voltage value contained in the state vector x , the deviation for each harmonic can be calculated from which the corresponding current value can be determined. In particular, the compensating current component is determined from the deviation, or alternatively from the detected harmonics of the voltage, respectively, i.e. for each harmonic. Such a single compensation current component or a plurality of such compensation current components are then superimposed and supplied together with the current value for the fundamental oscillation together with the reactive power if necessary.

The control block 416 then finally determines the current setting value i _s that is input to the inverter 410. These current settings take into account the current to be supplied in addition to superposition, i.e. fundamental oscillation and harmonic (s). Inverter 410 may generate a corresponding current value based on this current setting value i _s . Alternatively, it is possible to separately generate the fundamental oscillation of the current to be fed and at least one overlapping harmonic, and then superimpose, i.e. add, it may be provided. To this end, it is proposed to provide a plurality of inverters, at least one of which generates a fundamental oscillation or a part thereof, and at least the other generates at least one harmonic or a part thereof.

Therefore, the concept of the present invention has the following features in particular.

In modern energy systems, generating units and consuming devices are increasingly connected to a transmission network or distribution network through frequency converters or rectifiers. Due to the switching components of these power electrical systems, a sinusoidal 50 Hz or 60 Hz AC voltage is contaminated by harmonics in the frequency range up to about 10 kHz. These harmonics adversely affect the various operating means of the energy system, so if the level is too high, they are attenuated by the active or passive filter.

Wind turbines provide a significant portion of the energy already generated in some energy systems. Typically, a wind turbine is connected to the network via a frequency converter, and thus, in the worst case, it contributes to the harmonics of the network as well.

The invention described herein relates to a method for detecting and compensating for harmonics in a network voltage of an energy system by means of a wind turbine connected to the energy system via a full converter. Thus, the wind turbine (WEA) can supply current to reduce the voltage harmonic or reverse sequence system voltage. These may be generated, for example, by other generating units or consuming devices.

The following methods are particularly suggested for this purpose.

1. Measure terminal voltage of wind turbine of power plant transformer

2. Zero sequence of voltage Conversion of three terminal voltages to a rotating voltage indicator by Clarke-transformation under the adoption of the system's degree of freedom. [Voltage indicator is represented by amplitude and phase].

3. Application of a filter for monitoring the normal sequence system voltage and the reverse sequence system voltage of the DC component and a multiple of the commercial frequency up to the highest possible harmonic order defined in the primary, ie fundamental. Where the additional input to the filter is a differently determined commercial frequency. The filter may be formed by a state monitor, in particular a Kalman filter.

4. Calculation of the desired compensation current with the harmonic order selected for compensation via control law or control law. Preferably, an adjustment is made via a proportional constant (k _i ), where the proportional constant is i. Defines the generated compensating current of k _i A levels per V voltage harmonic amplitude for a voltage of order. Another desirable calculation of the compensating current is a calculation by application of a PID-adjuster to the difference between the set voltage, for example 0 V and the actual voltage of the corresponding component. Thus, with the sufficient current capability of the converter - ideally - a complete stabilization of the harmonic or reverse sequence system voltage can be achieved.

5. Set the compensation current in addition to the desired active current and reactive current in the basic vibration through the current control method. Preferably, a hysteresis current regulating device, also referred to as a tolerance band adjusting device, is used because a very fast response time and high robustness can be achieved here compared to the variable network parameters.

Particular attention should be paid to proper filtering of voltage indicators to quickly and accurately determine the correct voltage amplitudes over various frequency ranges, such as DC voltage, fundamental vibration, and harmonics.

Conventional methods for determining harmonic components in voltage are often based on regressively formulated discrete Fourier transforms of measured variables. They are characterized by high numerical robustness. However, the filter quality is guaranteed only if the sampling frequency is an integer multiple of the distance between two frequency components to be identified or the fundamental vibration frequency of the measurement signal. Since the commercial frequency in the energy system in general is subject to slight variations and also because of the applicability of the method in a system having different nominal frequencies (e.g. 50 Hz and 60 Hz), it is possible to use different commercial frequencies and a constant sampling rate The high filter quality in the filter is very important for maximum possible use.

Various solutions to these problems are known, but all have significant disadvantages. One possibility is to change the sampling rate of the algorithm to a function of the commercial frequency, as can be seen in [1]. However, this variation of the sampling rate is often undesirable because other algorithms where a fixed sampling rate is important, such as in the case of a regulator designed for a time-sharing type, can also be performed on the calculation unit. Another possibility is the execution of a phase adjustment loop or a frequency adjustment loop, as can be seen in Ref. [2]. However, such algorithms can be monitored for insufficient filter effects, particularly in the case of abnormal network conditions, in connection with phase adjustment loops, or in connection with frequency adjustment loops, especially in digital implementations where significant stability problems can be monitored when the sampling rate is low.

The third solution category is the Kalman filter, which is based on the state space model of the dynamics of the network voltage. Here, a filter for a constant frequency as can be seen in Reference [3] in which linear dynamics are disclosed, and a filter for a variable frequency, as can be seen in Reference [4] based on a nonlinear differential equation, Can be distinguished. The latter filter recursively computes an optimal monitor matrix under a given noise parameter, where a state space description based on the current state variables at each time step is used in the synthesis equation. In this case, due to corresponding appropriate parameterization, excellent filter characteristics can be obtained even in abnormal network conditions, such as where the commercial frequency varies and the voltage is asymmetric or heavily loaded by harmonics, Since it can be considered in the state space description of the system.

However, a disadvantage of the extended Kalman filter method for monitoring the state variables of nonlinear systems is the recursive computation of the monitor matrix. To do this, it is necessary to multiply the matrix of state space descriptions several times in each calculation step, and multiply several times with the vector of the measurement and state variables. If the filter dimension is high to account for multiple harmonic orders, this negatively affects the required computation time. In a limited computational capacity, execution within a limited cycle time by the sampling rate is no longer possible depending on the situation

On the other hand, with the adoption of constant commercial frequencies, solving the algebraic Riccati equations, assuming time invariant and linear system equations, as can be seen in Ref. [3], has the potential to precompute the monitor matrix for the Kalman filter do. According to at least one embodiment, for different commercial frequencies here, a precomputation of the monitor matrix is proposed.

According to reference [3], the following frequency dependence system of equations is used as a basis for deriving a monitor implementation.

는 알파 DC 전압 성분 및 베타 DC전압 성분,

는 i. 차수(즉, i = 1인 경우 기본 진동이고, i > 1인 경우 고조파 성분임)의 알파 정상 시퀀스 시스템 전압 성분 또는 베타 정상 시퀀스 시스템 전압 성분이고,

는 i. 차수의 알파 역상 시퀀스 시스템 전압 성분 또는 베타 역상 시퀀스 시스템 전압 성분이다.In these equations, k represents the current implementation phase, T _s is the sampling time of the discrete implementation of the algorithm, f is the commercial frequency, and i is the value up to the number of highest harmonic orders (N) The index is taken,

An alpha DC voltage component and a beta DC voltage component,

I. Is the alpha normal sequence system voltage component or the beta normal sequence system voltage component of order (i. E., The fundamental oscillation when i = 1 and the harmonic component when i> 1)

I. Order alpha-reversed sequence system voltage component or a beta reverse-phase sequence system voltage component.

Based on these equations, for each commercial frequency f, a monitor matrix K, which can be used to monitor the state at the time of a standard monitor implementation, can be calculated. The equation for calculating the Kalman monitor matrix using the state and measurement signal noise covariance matrices (Q and R) is known, for example, from reference [6].

Using the matrix K, an update is made to the state estimate at each time step according to known equations.

는 현재 시간 단계에서의 상태 추정을 나타내며,

는 최종 시간 단계에서 현재 시간 단계에 대한 상태 추정을 나타낸다. here

Represents the state estimation at the current time step,

Represents the state estimate for the current time step in the final time step.

In accordance with at least one embodiment, in accordance with at least one embodiment, in the method used herein, the monitor matrix comprises a dictionary of defined measurement and state noise levels for various commercial frequencies, including a predetermined state space description, . At runtime, the frequency is determined by an algorithm that is independent of the Kalman filter. This can be implemented by block frequency determination as shown in the figure. Here, for example, an appropriate filter can be used to determine the commercial frequency from the spatial indicator voltage angle. These frequency signals are used to select a monitor matrix suitable for these frequencies. Where interpolation may be performed to increase the resolution between the differently stored monitor matrices at different frequencies.

Thus, using this method, it is possible to accurately and quickly calculate multiple harmonics of the network voltage in a normal sequence system and a reversed phase sequence system with a greatly reduced computational cost compared to a nonlinear extended Kalman filter. Unlike the use of frequency control loops, for example, as described in [5], which also uses Kalman filters, the algorithms described here always represent a stability risk of the filter algorithm through nonlinear dynamics that are difficult to analyze , Do not use a closed loop to determine the commercial frequency. Rather, the commercial frequency is determined separately and is used only to select the monitor matrix.

The method described above can also be used to determine the harmonic of the current.

The above references are as follows.

[1] BP McGrath, DG Holmes, and J. Galloway: Improved Power Converter Line Synchronization Using Adaptive Discrete Fourier Transform (DFT). IEEE 33 ^rd Annual Power Electronics Expert Meeting Minutes, Vol. 2, 2002, pages 821-826.

[2] R. Teodorescu, M. Liserre and P. Rodriguez: Grid converters for photovoltaic and wind power systems, John Wiley & Sons, Chichester, 2011.

[3] R. R. Bitmead, A. C. Tsoi, P. J. Parker: Kalman filtering approach to short-time Fourier analysis, IEEE Transactions on Sound, Speech and Signal Processing, ASSP-34 (6), December 1986, pp. 1493-1501.

[4] B. La Scala, R. R. Bitmead: Design of an Extended Kalman Filter Frequency Tractor, Transactions on IEEE Signal Processing, Vol. 44 (3), March 1996, pp. 739-742.

[5] MS Reza, M. Ciobotaru, VG Agelidis: Instantaneous power quality analysis using frequency adaptive Kalman filter technique, The 7th International Power Electronics and Motion Control Conference minutes, Harbin, China, June 2012, pages 81-87 .

[6] T. Glad, L. Ljung: Control Theory, Taylor & Francis, London, 2000.

Claims (15)

A method for detecting a voltage in an electrical three-phase supply network according to fundamental oscillation and amplitude and phase for at least one harmonic,
- measuring the electrical three-phase voltage of the supply network;
- converting the measured voltage value into a polar coordinate comprising a rotation voltage indicator for said fundamental vibration as a measured reference indicator;
Monitoring, using a status monitor, at least one voltage indicator for said fundamental oscillation and at least one voltage indicator for at least one harmonic detected;
- updating said monitored values in accordance with said measured reference indicator
/ RTI > The method according to claim 1,
Wherein the status monitor operates according to a detected commercial frequency and the detected commercial frequency is input as an input variable to the status monitor. 3. The method according to claim 1 or 2,
Wherein the status monitor monitors the value of the voltage indicator using a current monitor matrix and wherein the current monitor matrix is selected from a plurality of pre-computed monitor matrices according to the detected commercial frequencies. 4. The method according to any one of claims 1 to 3,
Characterized in that a Kalman filter is used as the state monitor. 5. The method according to any one of claims 1 to 4,
Wherein a normal sequence system and a reverse sequence system are respectively detected for the fundamental vibration and each harmonic detected through the status monitor, and a DC component is selectively detected. 6. The method according to any one of claims 1 to 5,
The voltage indicator for the normal sequence system and the voltage indicator for the reverse phase sequence system are each based on the basic oscillation and each harmonic detected and their values are monitored and optionally the voltage indicator for the DC component is based, Is detected. 7. The method according to any one of claims 1 to 6,
The step of measuring the electrical three-phase voltage of the supply network is carried out in a transformer connected to the electrical three-phase supply network, and in particular electrically connected to the wind turbine supplying power to the electrical three-phase supply network via the transformer Characterized in that it is carried out on the side of the transformer facing it. 8. The method according to any one of claims 1 to 7,
Wherein the step of converting the measured voltage value to polar coordinates is performed using a Clarke-transformation under the adoption of a zero sequence system degree of freedom of voltage. A method for powering an electrical three-phase supply network,
Detecting a voltage in the supply network by a method according to any one of claims 1 to 8, wherein at least one harmonic is detected, and
- generating power for supplying to the electricity supply network, the power comprising a compensating current component to reduce at least one detected harmonic,
/ RTI > 10. The method of claim 9,
Wherein the compensating current component comprises at least one current harmonic and wherein the current harmonics are determined via amplitude and phase respectively and according to the monitored harmonics of the detected voltage. 11. The method according to claim 9 or 10,
The amplitudes of the current harmonics of the compensating current components are each adjusted via a regulating device and the regulating device is adapted to receive as an input signal a regulating difference between the voltage setpoint and the monitored voltage actual value of the corresponding harmonic of the detected voltage Lt; / RTI > 12. The method of claim 11,
As said regulating device,
- PID control device
- PI adjusters,
- P-adjusters and
- PD-regulator
≪ / RTI > wherein the adjusting device is selected from the list including the adjusting device. 13. The method according to any one of claims 9 to 12,
Wherein the compensation current component considers current harmonics as a normal sequence system component and a reverse phase sequence system component, respectively. 13. An inverter for generating power to be supplied to a three-phase power network, the inverter being prepared to operate in a method according to any one of claims 9 to 13. A wind turbine comprising an inverter according to claim 14 and / or being ready to carry out the method according to any one of claims 1 to 13.

Images